Event ·Official first-hand

NVIDIA 将 Run:ai GPU 编排平台以 Apache 2.0 协议开源

NVIDIA 把 Run:ai GPU 编排平台完整开源，采用 Apache 2.0 协议。发布内容包含 KAI Scheduler 调度器、CUDA 13.4 的 GPU 切分（fractioning）能力，以及配额管理功能。开发者还可以用本地 Kind 环境直接在 Kubernetes 上体验这套 GPU 资源编排工具。

Current read

NVIDIA 把 Run:ai 整个开源了，Apache 2.0 协议，K8s 上管 GPU 调度和切分的功能全放出来，做训练平台的可以上手试试。

11 sources71° AI heatLast updated 2026/10/19 03:19:18

Evidence

11 sources

Primary sourcemarktechpost

2026/10/19 03:19:18View source

Related source 1berryxia

2026/10/20 11:34:11View source

Related source 2kimmonismus

2026/10/20 10:26:39View source

Related source 3Teortaxes

2026/10/20 10:38:58View source

Related source 4AI21 Labs

2026/10/20 11:53:16View source

Related source 5@atomic_chat_hq

2026/10/20 12:42:02View source

Related source 6SuperTechFans

2026/10/21 00:16:33View source

Related source 7TestingCatalog

2026/10/17 10:49:40View source

Related source 8PyTorch

2026/10/17 16:00:51View source

Related source 9Gorden Sun

2026/10/18 14:13:49View source

Related source 10Thomas Wolf

2026/10/18 14:34:40View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story